Upload cfg.yaml

cfg.yaml

@@ -30,8 +30,7 @@ dataset:
     text_answer_separator: <|answer|>
     text_prompt_start: <|prompt|>
     text_system_start: <|system|>
-    train_dataframe: /app/h2o-llmstudio/data/user/Drug 5000/Drug Training - Training
-        Data AT.csv
+    train_dataframe: /app/h2o-llmstudio/data/user/Test_1_5000/output.csv
     validation_dataframe: None
     validation_size: 0.01
     validation_strategy: automatic
@@ -65,7 +64,7 @@ prediction:
     num_history: 4
     repetition_penalty: 1.2
     stop_tokens: ''
-    temperature: 0.3
+    temperature: 0.0
     top_k: 0
     top_p: 1.0
 problem_type: text_causal_language_modeling
@@ -81,7 +80,7 @@ training:
     adaptive_kl_control: true
     advantages_gamma: 0.99
     advantages_lambda: 0.95
-    batch_size: 8
+    batch_size: 6
     differential_learning_rate: 1.0e-05
     differential_learning_rate_layers: []
     drop_last_batch: true